Quantifying Systematic Age Discrepancies in Very Young Star Clusters

TL;DR

This study applies the Stellar Ages Bayesian framework to three young clusters to test age-tracer consistency. It finds a systematic offset between ages inferred from red supergiants (RSGs) and those from bright blue stars, with blue stars appearing ~0.55 dex younger (t_B ≈ t_RSG/3.55). It also reveals a population-level mismatch: RSGs overpredict the number of main-sequence stars by factors of ~3–5 when converted under single-star evolution, while blue stars imply much younger ages without corresponding low-mass MS counterparts. The results underscore the need to incorporate binaries and rapid rotation into stellar evolution models and advocate anchoring ages to RSGs while treating bright blue stars as a composite, non-single-star population; this has broad implications for massive-star evolution and UV-derived star-formation estimates.

Abstract

We infer the ages of three young stellar clusters, NGC 2004, NGC 7419, and NGC 2100, using Stellar Ages, a statistical algorithm designed to infer stellar population properties from color magnitude diagrams. Recent studies have revealed emerging inconsistencies in the inferred ages of very young stellar clusters with ages less than or equal to 50 Myr. Here, we identify and quantify two distinct discrepancies. First, we identify a systematic age offset of 0.55 plus minus 0.09 dex between red supergiant and bright blue star age estimates, equivalent to a factor of approximately 3.5 in linear age, with bright blue star ages appearing systematically younger than those inferred from red supergiants. Second, given the observed numbers of red supergiants and bright blue stars, we find a pronounced deficit of lower-mass main-sequence stars relative to expectations from a standard initial mass function. Although these discrepancies resemble those reported for intermediate-age clusters, their magnitude and character suggest that they are unique to the evolution of massive stars. Together, these results highlight population-level inconsistencies with single-star evolutionary models and underscore the need to consider multiple evolutionary tracers when age-dating young clusters. By combining individual stellar ages with population-wide constraints, our approach refines prior work on cluster age determinations and provides new insight into massive star evolution and the interpretation of cluster demographics.

Figures (8)

• Figure 1: Spatial distribution of RSG stars in the NGC 2100 field. Each of the 18 RSG points come from Table 1 in Beasor2016_NGC2100RSGs. The purple cross marks the field center (RA=85.4968°, DEC=-69.2083°), which is defined as the mean of the 18 coordinates. The red star shows the NGC 2100 cluster center defined by SIMBAD, and the dashed red circle indicates the 6 pc radius around that cluster center. $\sim$6 pc corresponds to two times the core radius of NGC 2100 defined in Table 1 of Niederhofer2015. Blue points represent RSGs outside the cluster radius (15 stars), while orange points show RSGs within the cluster (3 stars). The identifiers of the 3 nearby stars are compiled under Table \ref{['tab:RSG_subset']}.
• Figure 2: Color-Magnitude and Magnitude-Magnitude diagrams of cluster stars and appended RSG stars in NGC 2100. The cluster stars are from Niederhofer2015. Magnitude limits of 13.0 to 18.5 have been applied to both bands to retain stars with magnitude uncertainty of 0.1 or less. Top panel: B-V versus V magnitude with three RSG stars identified in Figure \ref{['fig:spatial_distribution']} highlighted in orange. The top panel presents a histogram across B-V, while the right panel presents a histogram across V magnitudes. Bottom panel: B versus V magnitude diagram with the three RSG stars indicated in orange.
• Figure 3: Inferred ages, metallicities, and rotation weights for NGC 2004 (a : left two columns), NGC 7419 (b: middle two columns), and NGC 2100 (c : right two columns), based on MCMC sampling with MIST single-star evolutionary models. Each violin summarizes the posterior distribution of weights, with the central dot representing the median. None of the clusters show a single dominant solution, underscoring the emerging inconsistencies. That said, some of the strongest posterior signals for all three clusters lie near $\log_{10}$(t/yr) $\sim$ 7.3–7.4 ($\sim$20–25 Myr), in good agreement with the RSG-based ages of $\sim$20–24 Myr reported by Beasor2019. These preferred solutions contrast with their younger MS and luminosity-function ages of $\sim$7–10 Myr, which we find less support for, suggesting that the stars driving these younger age estimates may not represent the bulk cluster population.
• Figure 4: Posterior distributions of age (top row) and initial rotation (bottom row) weights for NGC 2004 (a: left column), NGC 7419 (b: middle column), and NGC 2100 (c: right column), marginalized over the other parameters. The age distributions are broad, with only a modest peak near $\log_{10}$(t/yr) $\sim$ 7.3 ($\sim$20 Myr), and NGC 7419 shows some additional support at older ages. The rotation distributions are flat, with no preference for a specific initial value, indicating that rotation provides little leverage in constraining these clusters. These results are consistent with Beasor2019, who found relatively minor shifts in inferred cluster ages when including rotation.
• Figure 5: Color–Magnitude (top) and Magnitude–Magnitude (bottom) diagrams with the most likely $\log_{10}$ age for each star in NGC 2004 (a), NGC 7419 (b), and NGC 2100 (c). Across all three clusters, the $\sim$20 Myr solution (log age $\sim$ 7.3) is supported by the RSGs and a consistent MS population, while a younger $\sim$3 Myr signal (log age $\sim$ 6.5) arises from a handful of bright blue stars that lack corresponding lower-mass MS members.